FieldData structs implemented

src/hurricanedata/datareader.cu

@@ -1,60 +0,0 @@
-#include "datareader.h"
-
-#include <netcdf>
-
-using namespace std;
-using namespace netCDF;
-
-std::vector<float> readData(std::string path, std::string variableName) {
-    netCDF::NcFile data(path, netCDF::NcFile::read);
-
-    multimap<string, NcVar> vars = data.getVars();
-
-    NcVar var = vars.find(variableName)->second;   
-
-    int length = 1;
-    for (NcDim dim: var.getDims()) {
-        length *= dim.getSize();
-    }
-
-    vector<float> vec(length);
-
-    var.getVar(vec.data());
-
-    return vec;
-}
-
-std::pair<float*, size_t> loadDataToDevice(std::string path, std::string variableName) {
-    netCDF::NcFile data(path, netCDF::NcFile::read);
-
-    multimap<string, NcVar> vars = data.getVars();
-
-    NcVar var = vars.find(variableName)->second;   
-
-    int length = 1;
-    for (NcDim dim: var.getDims()) {
-        length *= dim.getSize();
-    }
-
-    // Store NetCDF variable in pinned memory on host
-    float *h_array;
-
-    cudaMallocHost(&h_array, sizeof(float)*length);
-
-    var.getVar(h_array);
-
-    // Copy data to device
-    float *d_array;
-
-    cudaError_t status = cudaMalloc(&d_array, sizeof(float)*length);
-    if (status != cudaSuccess)
-        cout << "Error allocating memory: " << status << "\n";
-
-    cudaMemcpyAsync(d_array, h_array, sizeof(float)*length, cudaMemcpyHostToDevice);
-
-    cudaDeviceSynchronize(); // Heavy hammer synchronisation // TODO: Use streams
-
-    cudaFreeHost(h_array);
-
-    return std::pair(d_array, length);
-}

src/hurricanedata/datareader.h

@@ -1,10 +0,0 @@
-#ifndef DATAREADER_H
-#define DATAREADER_H
-
-#include <vector>
-#include <string>
-
-std::vector<float> readData(std::string path, std::string variableName);
-std::pair<float*, size_t> loadDataToDevice(std::string path, std::string variableName);
-
-#endif //DATAREADER_H

src/hurricanedata/fielddata.cu

@@ -0,0 +1,12 @@
+#include "fielddata.h"
+
+// __device__ float getVal(
+//     const FieldMetadata &md,
+//     const FieldData &d,
+//     const size_t &timeInd,
+//     const size_t &lonInd,
+//     const size_t &latInd,
+//     const size_t &levInd
+// ) {
+//     return d.valArrays[0][timeInd]; 
+// }

src/hurricanedata/fielddata.h

@@ -0,0 +1,44 @@
+#ifndef FIELDDATA_H
+#define FIELDDATA_H
+
+#include <vector>
+
+struct FieldMetadata {
+    size_t widthSize; // Number of different longitudes
+    size_t heightSize; // Number of different latitudes
+    size_t depthSize; // Number of different levels
+
+    // lons is a managed Unified Memory array of size widthCount that indicates 
+    // that getVal(t, i, j, k) is a value with longitude of lons[i].
+    // The other such arrays are similarly defined.
+    double *lons; 
+    double *lats;
+    double *levs;
+};
+
+using FieldMetadata = FieldMetadata;
+
+struct FieldData {
+    static constexpr size_t FILESNUM = 2; // Number of files stored in a FieldData struct.
+
+    // An array of length FILESNUM storing pointers to 4D arrays stored in device memory.
+    float *valArrays[FILESNUM];
+
+    size_t timeSize; // Number of different times
+    // times is a managed Unified Memory array of size timeSize that indicates 
+    // that getVal(md, d, t, i, j, k) is a value at time times[t].
+    int *times;
+};
+
+using FieldData = FieldData;
+
+// __device__ float getVal(
+//     const FieldMetadata &md,
+//     const FieldData &d,
+//     const size_t &timeInd,
+//     const size_t &lonInd,
+//     const size_t &latInd,
+//     const size_t &levInd
+// );
+
+#endif //FIELDDATA_H

src/hurricanedata/gpubuffer.cu

@@ -0,0 +1,70 @@
+#include "gpubuffer.h"
+#include "fielddata.h"
+#include "gpubufferhelper.h"
+
+#include <netcdf>
+
+
+using namespace std;
+using namespace netCDF;
+
+GPUBuffer::GPUBuffer(std::string path, std::string variableName):
+path(path), variableName(variableName) {
+    NcFile data(path, NcFile::read);
+
+    multimap<string, NcVar> vars = data.getVars();
+
+    cudaMallocManaged(&fmd, sizeof(FieldMetadata));
+
+    readAndAllocateAxis<double>(&fmd->lons, &fmd->widthSize, vars.find("lon")->second);
+    readAndAllocateAxis<double>(&fmd->lats, &fmd->heightSize, vars.find("lat")->second);
+    readAndAllocateAxis<double>(&fmd->levs, &fmd->depthSize, vars.find("lev")->second);
+}
+
+
+
+FieldData GPUBuffer::nextFieldData() {
+    NcFile data(path, NcFile::read);
+
+    multimap<string, NcVar> vars = data.getVars();
+
+    FieldData fd;
+    size_t timeSize;
+    readAndAllocateAxis(&fd.times, &fd.timeSize, vars.find("time")->second);
+
+    NcVar var = vars.find(variableName)->second;   
+    
+    int length = 1;
+    for (NcDim dim: var.getDims()) {
+        length *= dim.getSize();
+    }
+
+    // Store NetCDF variable in pinned memory on host
+    float *h_array;
+
+    cudaMallocHost(&h_array, sizeof(float)*length);
+
+    var.getVar(h_array);
+
+    // Copy data to device
+    // float *d_array;
+
+    cudaError_t status = cudaMalloc(&fd.valArrays[0], sizeof(float)*length);
+    if (status != cudaSuccess)
+        cout << "Error allocating memory: " << status << "\n";
+
+    cudaMemcpyAsync(fd.valArrays[0], h_array, sizeof(float)*length, cudaMemcpyHostToDevice);
+
+    cudaDeviceSynchronize(); // Heavy hammer synchronisation // TODO: Use streams
+
+    cudaFreeHost(h_array);
+
+    return fd;
+}
+
+GPUBuffer::~GPUBuffer() {
+    cudaFree(fmd->lons);
+    cudaFree(fmd->lats);
+    cudaFree(fmd->levs);
+    cudaFree(fmd);
+}

src/hurricanedata/gpubuffer.h

@@ -0,0 +1,24 @@
+#ifndef GPUBUFFER_H
+#define GPUBUFFER_H
+
+#include "fielddata.h"
+
+#include <string>
+
+class GPUBuffer {
+public:
+    GPUBuffer(std::string path, std::string variableName);
+
+    FieldData nextFieldData();
+
+    ~GPUBuffer();
+
+    FieldMetadata *fmd;
+private:
+    // TODO: Implement GPUBuffer
+    std::string path;
+    std::string variableName;
+
+};
+
+#endif //GPUBUFFER_H

src/hurricanedata/gpubufferhelper.h

@@ -0,0 +1,10 @@
+#include <netcdf>
+#include <cassert>
+template <typename T>
+void readAndAllocateAxis(T** axis_ptr, size_t *size, const netCDF::NcVar &var) {
+    assert(var.getDimCount() == 1);
+    netCDF::NcDim dim = var.getDim(0);
+    *size = dim.getSize();
+    cudaError_t status = cudaMallocManaged(axis_ptr, *size*sizeof(T));
+    var.getVar(*axis_ptr);
+}

src/main.cu

@@ -1,19 +1,31 @@
-#include "hurricanedata/datareader.h"
+#include "hurricanedata/gpubuffer.h"
 
 #include <cuda_runtime.h>
 #include <device_launch_parameters.h>
 #include <iostream>
 #include <cmath>
 
+__device__ float getVal(
+    const FieldMetadata &md,
+    const FieldData &d,
+    const size_t &timeInd,
+    const size_t &lonInd,
+    const size_t &latInd,
+    const size_t &levInd
+) {
+    return d.valArrays[0][timeInd]; 
+}
+
 // Not parallel computation
-__global__ void computeMean(float *ans, size_t *masked_vals, size_t n, float *x) {
+__global__ void computeMean(float *ans, size_t *masked_vals, const FieldMetadata &fmd, FieldData fd) {
     float sum = 0;
     size_t num_not_masked_values = 0;
     size_t num_masked_values = 0;
-    for (int i = 0; i < n; i++) {
-        if (x[i] < 1E14) { /* If x is not missing value */
+    for (int i = 0; i < fmd.widthSize*fmd.heightSize*fmd.depthSize*fd.timeSize; i++) {
+        double xi = getVal(fmd, fd, i, 0, 0, 0);
+        if (xi < 1E14) { /* If x is not missing value */
             num_not_masked_values++;
-            sum += x[i];
+            sum += xi;
         } else {
             num_masked_values++;
         }
@@ -25,7 +37,9 @@ __global__ void computeMean(float *ans, size_t *masked_vals, size_t n, float *x)
 int main() {
     std::string path = "data/MERRA2_400.inst6_3d_ana_Np.20120101.nc4";
     std::string variable = "T";
-    auto arr = loadDataToDevice(path, variable);
+    GPUBuffer buffer{path, variable};
+
+    auto fd = buffer.nextFieldData();
 
     float *ptr_mean;
     cudaMallocManaged(&ptr_mean, sizeof(float));
@@ -33,15 +47,14 @@ int main() {
     size_t *ptr_masked;
     cudaMallocManaged(&ptr_masked, sizeof(size_t));
 
-    computeMean<<<1, 1>>>(ptr_mean, ptr_masked, arr.second, arr.first);
+    computeMean<<<1, 1>>>(ptr_mean, ptr_masked, *buffer.fmd, fd);
 
     cudaDeviceSynchronize();
 
-    std::cout << "Mean = " << *ptr_mean << " calculated from " << arr.second << " values where " << *ptr_masked << " are masked values.\n";
+    std::cout << "Mean = " << *ptr_mean << " values where " << *ptr_masked << " are masked values.\n";
 
-    cudaFree(arr.first);
+    cudaFree(fd.valArrays[0]);
     cudaFree(ptr_mean);
     cudaFree(ptr_masked);
-
     return 0;
 }

test_read.py

@@ -1,6 +1,5 @@
 import numpy as np
 from netCDF4 import Dataset
-from math import prod
 
 # Load the NetCDF file
 file_path = 'data/MERRA2_400.inst6_3d_ana_Np.20120101.nc4'
@@ -12,19 +11,23 @@ print(ncfile.variables.keys())
 U = ncfile.variables['T'][:] 
 
 # Check the shape of the variable 
-print(f"Shape of U: {U.shape} and total length is {prod(U.shape)}")
+print("Shape of U:", U.shape)
 
 # Compute the mean of the variable across all axes (for all elements in U)
 U_mean = np.mean(U)
+U_sum = np.sum(U)
 
 # Print the mean
 print("Mean of U:", U_mean)
 
-print(f"{U[0,0,0,1]=}")
-is_masked = np.ma.isMaskedArray(U)
-print(f"Is U a masked array? {is_masked}")
-masked_count = np.ma.count_masked(U)
-print("Number of masked values in U:", masked_count)
+print("Sum of U:", U_sum)
+
+sumval = 0
+row = U[0,0,100]
+for val in row:
+    if not np.ma.is_masked(val):
+        sumval += val
+print(f"Why does {np.sum(row)=} not equal {sumval=} ?!")
 
 # Close the NetCDF file
 ncfile.close()